Gersten Jonker

114   Chapter 5 ACTY students more frequently scored “meets expectations” than group C students. Ordinal logistic regression showed that group type had a statistically significant effect on the prediction of getting a high GRS on CBDs, OSCEs, and simulations (see Appendix 3 for a comprehensive report of the model). This finding can be rephrased with odds ratios of getting a high GRS, using group C as a reference (or: standard). The odds of participants in group ACTY getting a high GRS were 2.06 – 2.49 that in group C, statistically significant for OSCEs and simulations. Participants in group P had statistically significant higher odds (OR 2.45 – 6.69) of receiving a high GRS than those in group C. Also, the odds of obtaining a high GRS in group P was 2.70 for CBDs, 0.99 for OSCEs, and 3.23 for simulations compared to group ACTY; a statistically significant difference on simulations (Appendix 2). TABLE 3: Global rating scores (GRS) in post-test per assessment modality. Frequency of scores, per group type. Pie charts showing relative occurrence of global rating scores on CBDs, OSCEs, simulations. Group ACTY n (%) Group C n (%) Group P n (%) GRS on CBDs Does not meet expectations 2 (3.4) 4 (11.1) 0 (0) Borderline 7 (12.1) 7 (19.4) 2 (6.5) Meets expectations 49 (84.5) 25 (69.4) 29 (93.5) Pie chart (%) GRS on OSCEs Does not meet expectations 12 (7.4) 12 (13.2) 4 (4.9) Borderline 25 (15.3) 27 (29.7) 15 (18.5) Meets expectations 126 (77.3) 52 (57.1) 62 (76.5) Pie chart (%) GRS on Simulations Does not meet expectations 18 (22.5) 14 (33.3) 4 (7.8) Borderline 25 (31.3) 17 (40.5) 10 (19.6) Meets expectations 37 (46.3) 11 (26.2) 37 (72.5) Pie chart (%) ACTY= Acute Care Transitional Year, C= student comparison group, P= PNIT comparison group. CBD= case-based discussion, OSCE= objective structured clinical examination. Because each participant did more than one station per modality, n refers to the times that, on a modality, the respective score was awarded.